{"id":1157080,"date":"2025-01-13T18:24:04","date_gmt":"2025-01-13T10:24:04","guid":{"rendered":"https:\/\/docs.pingcode.com\/ask\/ask-ask\/1157080.html"},"modified":"2025-01-13T18:24:07","modified_gmt":"2025-01-13T10:24:07","slug":"python%e7%a8%8b%e5%ba%8f%e5%a6%82%e4%bd%95%e5%8a%a0%e5%85%a5%e5%99%aa%e5%a3%b0","status":"publish","type":"post","link":"https:\/\/docs.pingcode.com\/ask\/ask-ask\/1157080.html","title":{"rendered":"python\u7a0b\u5e8f\u5982\u4f55\u52a0\u5165\u566a\u58f0"},"content":{"rendered":"

\"python\u7a0b\u5e8f\u5982\u4f55\u52a0\u5165\u566a\u58f0\"<\/p>\n

\u5728Python\u7a0b\u5e8f\u4e2d\u52a0\u5165\u566a\u58f0\u7684\u65b9\u6cd5\u6709\u591a\u79cd\uff0c\u53ef\u4ee5\u901a\u8fc7\u4f7f\u7528NumPy\u5e93\u751f\u6210\u968f\u673a\u566a\u58f0\u3001\u5728\u6570\u636e\u96c6\u4e2d\u6dfb\u52a0\u9ad8\u65af\u566a\u58f0\u3001\u76d0\u548c\u80e1\u6912\u566a\u58f0\u7b49<\/strong>\u3002\u5176\u4e2d\uff0cNumPy\u5e93\u662f\u6700\u5e38\u7528\u7684\u5de5\u5177\u4e4b\u4e00\uff0c\u56e0\u4e3a\u5b83\u63d0\u4f9b\u4e86\u4e30\u5bcc\u7684\u968f\u673a\u6570\u751f\u6210\u51fd\u6570\uff0c\u53ef\u4ee5\u65b9\u4fbf\u5730\u751f\u6210\u5404\u79cd\u7c7b\u578b\u7684\u566a\u58f0\u3002\u63a5\u4e0b\u6765\uff0c\u6211\u4eec\u5c06\u8be6\u7ec6\u4ecb\u7ecd\u5982\u4f55\u4f7f\u7528NumPy\u5e93\u751f\u6210\u968f\u673a\u566a\u58f0\uff0c\u5e76\u5728\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0\u3002<\/p>\n<\/p>\n

\u4e00\u3001\u4f7f\u7528NumPy\u751f\u6210\u968f\u673a\u566a\u58f0<\/h2>\n<\/p>\n

NumPy\u5e93\u662f\u4e00\u4e2a\u5f3a\u5927\u7684\u79d1\u5b66\u8ba1\u7b97\u5e93\uff0c\u63d0\u4f9b\u4e86\u591a\u79cd\u751f\u6210\u968f\u673a\u6570\u7684\u65b9\u6cd5\u3002\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u5b83\u6765\u751f\u6210\u5404\u79cd\u7c7b\u578b\u7684\u566a\u58f0\uff0c\u5982\u9ad8\u65af\u566a\u58f0\u3001\u5747\u5300\u566a\u58f0\u7b49\u3002<\/p>\n<\/p>\n

1\u3001\u751f\u6210\u9ad8\u65af\u566a\u58f0<\/h3>\n<\/p>\n

\u9ad8\u65af\u566a\u58f0\uff0c\u4e5f\u79f0\u4e3a\u6b63\u6001\u566a\u58f0\uff0c\u662f\u6570\u636e\u79d1\u5b66\u548c\u673a\u5668\u5b66\u4e60<\/a>\u4e2d\u5e38\u7528\u7684\u4e00\u79cd\u566a\u58f0\u7c7b\u578b\u3002\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528NumPy\u7684numpy.random.normal<\/code>\u51fd\u6570\u751f\u6210\u9ad8\u65af\u566a\u58f0\u3002<\/p>\n<\/p>\n

import numpy as np<\/p>\n
\u751f\u6210\u9ad8\u65af\u566a\u58f0<\/strong><\/h2>\n
mean = 0  # \u5747\u503c<\/p>\n
std_dev = 1  # \u6807\u51c6\u5dee<\/p>\n
num_samples = 1000  # \u6837\u672c\u6570\u91cf<\/p>\n
gaussian_noise = np.random.normal(mean, std_dev, num_samples)<\/p>\n
print(gaussian_noise)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u4e0a\u8ff0\u4ee3\u7801\u4e2d\uff0c\u6211\u4eec\u751f\u6210\u4e861000\u4e2a\u5747\u503c\u4e3a0\uff0c\u6807\u51c6\u5dee\u4e3a1\u7684\u9ad8\u65af\u566a\u58f0\u6837\u672c\u3002<\/p>\n<\/p>\n
2\u3001\u751f\u6210\u5747\u5300\u566a\u58f0<\/h3>\n<\/p>\n
\u5747\u5300\u566a\u58f0\u662f\u4e00\u79cd\u5206\u5e03\u5728\u4e00\u5b9a\u533a\u95f4\u5185\u7684\u968f\u673a\u566a\u58f0\u3002\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528numpy.random.uniform<\/code>\u51fd\u6570\u751f\u6210\u5747\u5300\u566a\u58f0\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
\u751f\u6210\u5747\u5300\u566a\u58f0<\/strong><\/h2>\n
low = -1  # \u4e0b\u9650<\/p>\n
high = 1  # \u4e0a\u9650<\/p>\n
num_samples = 1000  # \u6837\u672c\u6570\u91cf<\/p>\n
uniform_noise = np.random.uniform(low, high, num_samples)<\/p>\n
print(uniform_noise)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u4e0a\u8ff0\u4ee3\u7801\u4e2d\uff0c\u6211\u4eec\u751f\u6210\u4e861000\u4e2a\u5728\u533a\u95f4[-1, 1]\u5185\u5747\u5300\u5206\u5e03\u7684\u566a\u58f0\u6837\u672c\u3002<\/p>\n<\/p>\n
\u4e8c\u3001\u5728\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0<\/h2>\n<\/p>\n
\u5728\u751f\u6210\u566a\u58f0\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u5c06\u5176\u6dfb\u52a0\u5230\u6570\u636e\u96c6\u4e2d\uff0c\u4ee5\u589e\u52a0\u6570\u636e\u7684\u591a\u6837\u6027\uff0c\u63d0\u9ad8\u6a21\u578b\u7684\u9c81\u68d2\u6027\u3002\u4ee5\u4e0b\u662f\u51e0\u79cd\u5e38\u89c1\u7684\u6570\u636e\u96c6\u52a0\u566a\u65b9\u6cd5\uff1a<\/p>\n<\/p>\n
1\u3001\u5728\u56fe\u50cf\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0<\/h3>\n<\/p>\n
\u56fe\u50cf\u5904\u7406\u662f\u6570\u636e\u79d1\u5b66\u4e2d\u7684\u4e00\u4e2a\u91cd\u8981\u9886\u57df\u3002\u5728\u56fe\u50cf\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u589e\u52a0\u6a21\u578b\u7684\u9c81\u68d2\u6027\u548c\u6cdb\u5316\u80fd\u529b\u3002<\/p>\n<\/p>\n
1.1\u3001\u6dfb\u52a0\u9ad8\u65af\u566a\u58f0<\/h4>\n<\/p>\n
import numpy as np<\/p>\n
import cv2<\/p>\n
import matplotlib.pyplot as plt<\/p>\n
\u8bfb\u53d6\u56fe\u50cf<\/strong><\/h2>\n
image = cv2.imread('image.jpg', cv2.IMREAD_GRAYSCALE)<\/p>\n
\u751f\u6210\u9ad8\u65af\u566a\u58f0<\/strong><\/h2>\n
mean = 0<\/p>\n
std_dev = 25<\/p>\n
gaussian_noise = np.random.normal(mean, std_dev, image.shape)<\/p>\n
\u5c06\u566a\u58f0\u6dfb\u52a0\u5230\u56fe\u50cf<\/strong><\/h2>\n
noisy_image = image + gaussian_noise<\/p>\n
\u663e\u793a\u539f\u56fe\u548c\u52a0\u566a\u56fe\u50cf<\/strong><\/h2>\n
plt.figure(figsize=(10, 5))<\/p>\n
plt.subplot(1, 2, 1)<\/p>\n
plt.title('Original Image')<\/p>\n
plt.imshow(image, cmap='gray')<\/p>\n
plt.subplot(1, 2, 2)<\/p>\n
plt.title('Noisy Image')<\/p>\n
plt.imshow(noisy_image, cmap='gray')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
1.2\u3001\u6dfb\u52a0\u76d0\u548c\u80e1\u6912\u566a\u58f0<\/h4>\n<\/p>\n
\u76d0\u548c\u80e1\u6912\u566a\u58f0\u662f\u4e00\u79cd\u968f\u673a\u5c06\u50cf\u7d20\u503c\u8bbe\u4e3a\u6700\u5927\u503c\u6216\u6700\u5c0f\u503c\u7684\u566a\u58f0\u7c7b\u578b\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
import cv2<\/p>\n
import random<\/p>\n
import matplotlib.pyplot as plt<\/p>\n
\u8bfb\u53d6\u56fe\u50cf<\/strong><\/h2>\n
image = cv2.imread('image.jpg', cv2.IMREAD_GRAYSCALE)<\/p>\n
\u6dfb\u52a0\u76d0\u548c\u80e1\u6912\u566a\u58f0<\/strong><\/h2>\n
def add_salt_and_pepper_noise(image, prob):<\/p>\n
    noisy_image = np.copy(image)<\/p>\n
    num_salt = np.ceil(prob * image.size * 0.5)<\/p>\n
    num_pepper = np.ceil(prob * image.size * 0.5)<\/p>\n
    # \u6dfb\u52a0\u76d0\u566a\u58f0<\/p>\n
    coords = [np.random.randint(0, i - 1, int(num_salt)) for i in image.shape]<\/p>\n
    noisy_image[coords[0], coords[1]] = 255<\/p>\n
    # \u6dfb\u52a0\u80e1\u6912\u566a\u58f0<\/p>\n
    coords = [np.random.randint(0, i - 1, int(num_pepper)) for i in image.shape]<\/p>\n
    noisy_image[coords[0], coords[1]] = 0<\/p>\n
    return noisy_image<\/p>\n
prob = 0.02<\/p>\n
noisy_image = add_salt_and_pepper_noise(image, prob)<\/p>\n
\u663e\u793a\u539f\u56fe\u548c\u52a0\u566a\u56fe\u50cf<\/strong><\/h2>\n
plt.figure(figsize=(10, 5))<\/p>\n
plt.subplot(1, 2, 1)<\/p>\n
plt.title('Original Image')<\/p>\n
plt.imshow(image, cmap='gray')<\/p>\n
plt.subplot(1, 2, 2)<\/p>\n
plt.title('Noisy Image')<\/p>\n
plt.imshow(noisy_image, cmap='gray')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
2\u3001\u5728\u65f6\u95f4\u5e8f\u5217\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0<\/h3>\n<\/p>\n
\u65f6\u95f4\u5e8f\u5217\u6570\u636e\u96c6\u5728\u91d1\u878d\u3001\u6c14\u8c61\u7b49\u9886\u57df\u4e2d\u5e7f\u6cdb\u5e94\u7528\u3002\u5728\u65f6\u95f4\u5e8f\u5217\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u63d0\u9ad8\u9884\u6d4b\u6a21\u578b\u7684\u9c81\u68d2\u6027\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
import matplotlib.pyplot as plt<\/p>\n
\u751f\u6210\u65f6\u95f4\u5e8f\u5217\u6570\u636e<\/strong><\/h2>\n
time = np.arange(0, 100, 0.1)<\/p>\n
data = np.sin(time)<\/p>\n
\u751f\u6210\u9ad8\u65af\u566a\u58f0<\/strong><\/h2>\n
mean = 0<\/p>\n
std_dev = 0.1<\/p>\n
gaussian_noise = np.random.normal(mean, std_dev, len(time))<\/p>\n
\u5c06\u566a\u58f0\u6dfb\u52a0\u5230\u65f6\u95f4\u5e8f\u5217\u6570\u636e<\/strong><\/h2>\n
noisy_data = data + gaussian_noise<\/p>\n
\u663e\u793a\u539f\u6570\u636e\u548c\u52a0\u566a\u6570\u636e<\/strong><\/h2>\n
plt.figure(figsize=(10, 5))<\/p>\n
plt.plot(time, data, label='Original Data')<\/p>\n
plt.plot(time, noisy_data, label='Noisy Data')<\/p>\n
plt.legend()<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
3\u3001\u5728\u6587\u672c\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0<\/h3>\n<\/p>\n
\u6587\u672c\u6570\u636e\u5904\u7406\u662f\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff08NLP\uff09\u4e2d\u7684\u4e00\u4e2a\u91cd\u8981\u73af\u8282\u3002\u5728\u6587\u672c\u6570\u636e\u96c6\u4e2d\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u589e\u52a0\u6a21\u578b\u7684\u9c81\u68d2\u6027\u548c\u6cdb\u5316\u80fd\u529b\u3002<\/p>\n<\/p>\n
3.1\u3001\u6dfb\u52a0\u5b57\u7b26\u566a\u58f0<\/h4>\n<\/p>\n
import random<\/p>\n
\u5b9a\u4e49\u51fd\u6570\u6dfb\u52a0\u5b57\u7b26\u566a\u58f0<\/strong><\/h2>\n
def add_char_noise(text, prob):<\/p>\n
    noisy_text = ''<\/p>\n
    for char in text:<\/p>\n
        if random.random() < prob:<\/p>\n
            noisy_text += chr(random.randint(32, 126))<\/p>\n
        else:<\/p>\n
            noisy_text += char<\/p>\n
    return noisy_text<\/p>\n
\u793a\u4f8b\u6587\u672c<\/strong><\/h2>\n
text = "Hello, World!"<\/p>\n
prob = 0.1<\/p>\n
noisy_text = add_char_noise(text, prob)<\/p>\n
print("Original Text:", text)<\/p>\n
print("Noisy Text:", noisy_text)<\/p>\n
<\/code><\/pre>\n<\/p>\n
3.2\u3001\u6dfb\u52a0\u5355\u8bcd\u566a\u58f0<\/h4>\n<\/p>\n
import random<\/p>\n
\u5b9a\u4e49\u51fd\u6570\u6dfb\u52a0\u5355\u8bcd\u566a\u58f0<\/strong><\/h2>\n
def add_word_noise(text, prob):<\/p>\n
    words = text.split()<\/p>\n
    noisy_words = []<\/p>\n
    for word in words:<\/p>\n
        if random.random() < prob:<\/p>\n
            noisy_words.append(''.join(random.sample(word, len(word))))<\/p>\n
        else:<\/p>\n
            noisy_words.append(word)<\/p>\n
    return ' '.join(noisy_words)<\/p>\n
\u793a\u4f8b\u6587\u672c<\/strong><\/h2>\n
text = "Hello, World! This is a test sentence."<\/p>\n
prob = 0.2<\/p>\n
noisy_text = add_word_noise(text, prob)<\/p>\n
print("Original Text:", text)<\/p>\n
print("Noisy Text:", noisy_text)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u4e09\u3001\u5728\u4e0d\u540c\u5e94\u7528\u573a\u666f\u4e2d\u52a0\u5165\u566a\u58f0<\/h2>\n<\/p>\n
\u4e0d\u540c\u7684\u5e94\u7528\u573a\u666f\u5bf9\u566a\u58f0\u7684\u9700\u6c42\u4e0d\u540c\uff0c\u52a0\u5165\u566a\u58f0\u7684\u65b9\u5f0f\u4e5f\u6709\u6240\u4e0d\u540c\u3002\u63a5\u4e0b\u6765\uff0c\u6211\u4eec\u5c06\u4ecb\u7ecd\u51e0\u4e2a\u5e38\u89c1\u7684\u5e94\u7528\u573a\u666f\uff0c\u5e76\u63a2\u8ba8\u5982\u4f55\u5728\u8fd9\u4e9b\u573a\u666f\u4e2d\u52a0\u5165\u566a\u58f0\u3002<\/p>\n<\/p>\n
1\u3001\u673a\u5668\u5b66\u4e60\u4e2d\u7684\u6570\u636e\u589e\u5f3a<\/h3>\n<\/p>\n
\u6570\u636e\u589e\u5f3a\u662f\u673a\u5668\u5b66\u4e60\u4e2d\u5e38\u7528\u7684\u6280\u672f\uff0c\u901a\u8fc7\u5bf9\u539f\u59cb\u6570\u636e\u8fdb\u884c\u53d8\u6362\uff0c\u751f\u6210\u65b0\u7684\u6570\u636e\u6837\u672c\uff0c\u4ee5\u63d0\u9ad8\u6a21\u578b\u7684\u6cdb\u5316\u80fd\u529b\u3002\u52a0\u5165\u566a\u58f0\u662f\u6570\u636e\u589e\u5f3a\u7684\u4e00\u79cd\u5e38\u89c1\u65b9\u6cd5\u3002<\/p>\n<\/p>\n
1.1\u3001\u56fe\u50cf\u5206\u7c7b\u4e2d\u7684\u6570\u636e\u589e\u5f3a<\/h4>\n<\/p>\n
\u5728\u56fe\u50cf\u5206\u7c7b\u4efb\u52a1\u4e2d\uff0c\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u589e\u52a0\u6570\u636e\u7684\u591a\u6837\u6027\uff0c\u63d0\u9ad8\u6a21\u578b\u7684\u9c81\u68d2\u6027\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
import cv2<\/p>\n
\u8bfb\u53d6\u56fe\u50cf<\/strong><\/h2>\n
image = cv2.imread('image.jpg', cv2.IMREAD_GRAYSCALE)<\/p>\n
\u751f\u6210\u9ad8\u65af\u566a\u58f0<\/strong><\/h2>\n
mean = 0<\/p>\n
std_dev = 25<\/p>\n
gaussian_noise = np.random.normal(mean, std_dev, image.shape)<\/p>\n
\u5c06\u566a\u58f0\u6dfb\u52a0\u5230\u56fe\u50cf<\/strong><\/h2>\n
noisy_image = image + gaussian_noise<\/p>\n
\u4fdd\u5b58\u52a0\u566a\u56fe\u50cf<\/strong><\/h2>\n
cv2.imwrite('noisy_image.jpg', noisy_image)<\/p>\n
<\/code><\/pre>\n<\/p>\n
1.2\u3001\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u4e2d\u7684\u6570\u636e\u589e\u5f3a<\/h4>\n<\/p>\n
\u5728\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u4efb\u52a1\u4e2d\uff0c\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u589e\u52a0\u6570\u636e\u7684\u591a\u6837\u6027\uff0c\u63d0\u9ad8\u6a21\u578b\u7684\u9c81\u68d2\u6027\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
import pandas as pd<\/p>\n
\u751f\u6210\u65f6\u95f4\u5e8f\u5217\u6570\u636e<\/strong><\/h2>\n
time = pd.date_range(start='1\/1\/2020', periods=100, freq='D')<\/p>\n
data = np.sin(np.linspace(0, 10, 100))<\/p>\n
\u751f\u6210\u9ad8\u65af\u566a\u58f0<\/strong><\/h2>\n
mean = 0<\/p>\n
std_dev = 0.1<\/p>\n
gaussian_noise = np.random.normal(mean, std_dev, len(time))<\/p>\n
\u5c06\u566a\u58f0\u6dfb\u52a0\u5230\u65f6\u95f4\u5e8f\u5217\u6570\u636e<\/strong><\/h2>\n
noisy_data = data + gaussian_noise<\/p>\n
\u521b\u5efaDataFrame<\/strong><\/h2>\n
df = pd.DataFrame({'Date': time, 'Value': noisy_data})<\/p>\n
print(df)<\/p>\n
<\/code><\/pre>\n<\/p>\n
2\u3001\u8bed\u97f3\u8bc6\u522b\u4e2d\u7684\u6570\u636e\u589e\u5f3a<\/h3>\n<\/p>\n
\u5728\u8bed\u97f3\u8bc6\u522b\u4efb\u52a1\u4e2d\uff0c\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u589e\u52a0\u6570\u636e\u7684\u591a\u6837\u6027\uff0c\u63d0\u9ad8\u6a21\u578b\u7684\u9c81\u68d2\u6027\u3002<\/p>\n<\/p>\n
2.1\u3001\u6dfb\u52a0\u767d\u566a\u58f0<\/h4>\n<\/p>\n
\u767d\u566a\u58f0\u662f\u4e00\u79cd\u5e38\u89c1\u7684\u566a\u58f0\u7c7b\u578b\uff0c\u53ef\u4ee5\u7528\u4e8e\u8bed\u97f3\u8bc6\u522b\u4e2d\u7684\u6570\u636e\u589e\u5f3a\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
import librosa<\/p>\n
import soundfile as sf<\/p>\n
\u8bfb\u53d6\u97f3\u9891\u6587\u4ef6<\/strong><\/h2>\n
audio, sr = librosa.load('audio.wav', sr=None)<\/p>\n
\u751f\u6210\u767d\u566a\u58f0<\/strong><\/h2>\n
mean = 0<\/p>\n
std_dev = 0.05<\/p>\n
white_noise = np.random.normal(mean, std_dev, len(audio))<\/p>\n
\u5c06\u566a\u58f0\u6dfb\u52a0\u5230\u97f3\u9891<\/strong><\/h2>\n
noisy_audio = audio + white_noise<\/p>\n
\u4fdd\u5b58\u52a0\u566a\u97f3\u9891\u6587\u4ef6<\/strong><\/h2>\n
sf.write('noisy_audio.wav', noisy_audio, sr)<\/p>\n
<\/code><\/pre>\n<\/p>\n
2.2\u3001\u6dfb\u52a0\u73af\u5883\u566a\u58f0<\/h4>\n<\/p>\n
\u73af\u5883\u566a\u58f0\u662f\u8bed\u97f3\u8bc6\u522b\u4e2d\u7684\u4e00\u79cd\u5e38\u89c1\u566a\u58f0\u7c7b\u578b\uff0c\u53ef\u4ee5\u901a\u8fc7\u5f55\u5236\u771f\u5b9e\u73af\u5883\u4e2d\u7684\u566a\u58f0\u6216\u4ece\u7f51\u7edc\u4e0a\u4e0b\u8f7d\u566a\u58f0\u6570\u636e\u6765\u5b9e\u73b0\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
import librosa<\/p>\n
import soundfile as sf<\/p>\n
\u8bfb\u53d6\u97f3\u9891\u6587\u4ef6<\/strong><\/h2>\n
audio, sr = librosa.load('audio.wav', sr=None)<\/p>\n
\u8bfb\u53d6\u73af\u5883\u566a\u58f0\u6587\u4ef6<\/strong><\/h2>\n
noise, sr_noise = librosa.load('noise.wav', sr=sr)<\/p>\n
\u5c06\u566a\u58f0\u6dfb\u52a0\u5230\u97f3\u9891<\/strong><\/h2>\n
noisy_audio = audio + noise[:len(audio)]<\/p>\n
\u4fdd\u5b58\u52a0\u566a\u97f3\u9891\u6587\u4ef6<\/strong><\/h2>\n
sf.write('noisy_audio.wav', noisy_audio, sr)<\/p>\n
<\/code><\/pre>\n<\/p>\n
3\u3001\u6587\u672c\u5206\u7c7b\u4e2d\u7684\u6570\u636e\u589e\u5f3a<\/h3>\n<\/p>\n
\u5728\u6587\u672c\u5206\u7c7b\u4efb\u52a1\u4e2d\uff0c\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u589e\u52a0\u6570\u636e\u7684\u591a\u6837\u6027\uff0c\u63d0\u9ad8\u6a21\u578b\u7684\u9c81\u68d2\u6027\u3002<\/p>\n<\/p>\n
3.1\u3001\u6dfb\u52a0\u62fc\u5199\u9519\u8bef<\/h4>\n<\/p>\n
\u62fc\u5199\u9519\u8bef\u662f\u4e00\u79cd\u5e38\u89c1\u7684\u6587\u672c\u566a\u58f0\u7c7b\u578b\uff0c\u53ef\u4ee5\u901a\u8fc7\u968f\u673a\u66ff\u6362\u3001\u5220\u9664\u6216\u63d2\u5165\u5b57\u7b26\u6765\u5b9e\u73b0\u3002<\/p>\n<\/p>\n
import random<\/p>\n
\u5b9a\u4e49\u51fd\u6570\u6dfb\u52a0\u62fc\u5199\u9519\u8bef<\/strong><\/h2>\n
def add_typo(text, prob):<\/p>\n
    noisy_text = ''<\/p>\n
    for char in text:<\/p>\n
        if random.random() < prob:<\/p>\n
            if random.random() < 0.5:<\/p>\n
                noisy_text += chr(random.randint(32, 126))<\/p>\n
            else:<\/p>\n
                noisy_text += ''<\/p>\n
        else:<\/p>\n
            noisy_text += char<\/p>\n
    return noisy_text<\/p>\n
\u793a\u4f8b\u6587\u672c<\/strong><\/h2>\n
text = "Hello, World!"<\/p>\n
prob = 0.1<\/p>\n
noisy_text = add_typo(text, prob)<\/p>\n
print("Original Text:", text)<\/p>\n
print("Noisy Text:", noisy_text)<\/p>\n
<\/code><\/pre>\n<\/p>\n
3.2\u3001\u6dfb\u52a0\u540c\u4e49\u8bcd\u66ff\u6362<\/h4>\n<\/p>\n
\u540c\u4e49\u8bcd\u66ff\u6362\u662f\u4e00\u79cd\u5e38\u89c1\u7684\u6587\u672c\u566a\u58f0\u7c7b\u578b\uff0c\u53ef\u4ee5\u901a\u8fc7\u5c06\u5355\u8bcd\u66ff\u6362\u4e3a\u5176\u540c\u4e49\u8bcd\u6765\u5b9e\u73b0\u3002<\/p>\n<\/p>\n
import random<\/p>\n
from nltk.corpus import wordnet<\/p>\n
\u5b9a\u4e49\u51fd\u6570\u6dfb\u52a0\u540c\u4e49\u8bcd\u66ff\u6362<\/strong><\/h2>\n
def add_synonym_noise(text, prob):<\/p>\n
    words = text.split()<\/p>\n
    noisy_words = []<\/p>\n
    for word in words:<\/p>\n
        if random.random() < prob:<\/p>\n
            synonyms = wordnet.synsets(word)<\/p>\n
            if synonyms:<\/p>\n
                synonym = synonyms[0].lemmas()[0].name()<\/p>\n
                noisy_words.append(synonym)<\/p>\n
            else:<\/p>\n
                noisy_words.append(word)<\/p>\n
        else:<\/p>\n
            noisy_words.append(word)<\/p>\n
    return ' '.join(noisy_words)<\/p>\n
\u793a\u4f8b\u6587\u672c<\/strong><\/h2>\n
text = "Hello, World! This is a test sentence."<\/p>\n
prob = 0.2<\/p>\n
noisy_text = add_synonym_noise(text, prob)<\/p>\n
print("Original Text:", text)<\/p>\n
print("Noisy Text:", noisy_text)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u56db\u3001\u52a0\u5165\u566a\u58f0\u7684\u4f18\u52bf\u548c\u6311\u6218<\/h2>\n<\/p>\n
1\u3001\u4f18\u52bf<\/h3>\n<\/p>\n
1.1\u3001\u63d0\u9ad8\u6a21\u578b\u7684\u9c81\u68d2\u6027<\/h4>\n<\/p>\n
\u52a0\u5165\u566a\u58f0\u53ef\u4ee5\u63d0\u9ad8\u6a21\u578b\u5728\u9762\u5bf9\u4e0d\u540c\u6570\u636e\u6837\u672c\u65f6\u7684\u9c81\u68d2\u6027\uff0c\u4f7f\u6a21\u578b\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u66f4\u5177\u7a33\u5b9a\u6027\u548c\u53ef\u9760\u6027\u3002<\/p>\n<\/p>\n
1.2\u3001\u589e\u52a0\u6570\u636e\u7684\u591a\u6837\u6027<\/h4>\n<\/p>\n
\u901a\u8fc7\u52a0\u5165\u566a\u58f0\uff0c\u53ef\u4ee5\u751f\u6210\u66f4\u591a\u6837\u5316\u7684\u6570\u636e\u6837\u672c\uff0c\u4ece\u800c\u63d0\u9ad8\u6a21\u578b\u7684\u6cdb\u5316\u80fd\u529b\uff0c\u51cf\u5c11\u8fc7\u62df\u5408\u73b0\u8c61\u3002<\/p>\n<\/p>\n
2\u3001\u6311\u6218<\/h3>\n<\/p>\n
2.1\u3001\u63a7\u5236\u566a\u58f0\u5f3a\u5ea6<\/h4>\n<\/p>\n
\u5728\u52a0\u5165\u566a\u58f0\u65f6\uff0c\u9700\u8981\u5408\u7406\u63a7\u5236\u566a\u58f0\u7684\u5f3a\u5ea6\uff0c\u4ee5\u907f\u514d\u5bf9\u539f\u59cb\u6570\u636e\u9020\u6210\u8fc7\u5927\u7684\u5e72\u6270\uff0c\u4ece\u800c\u5f71\u54cd\u6a21\u578b\u7684\u6027\u80fd\u3002<\/p>\n<\/p>\n
2.2\u3001\u9009\u62e9\u5408\u9002\u7684\u566a\u58f0\u7c7b\u578b<\/h4>\n<\/p>\n
\u4e0d\u540c\u7684\u5e94\u7528\u573a\u666f\u5bf9\u566a\u58f0\u7684\u9700\u6c42\u4e0d\u540c\uff0c\u9700\u8981\u6839\u636e\u5177\u4f53\u60c5\u51b5\u9009\u62e9\u5408\u9002\u7684\u566a\u58f0\u7c7b\u578b\uff0c\u4ee5\u5b9e\u73b0\u6700\u4f73\u7684\u589e\u5f3a\u6548\u679c\u3002<\/p>\n<\/p>\n
\u4e94\u3001\u603b\u7ed3<\/h2>\n<\/p>\n
\u5728Python\u7a0b\u5e8f\u4e2d\u52a0\u5165\u566a\u58f0\u662f\u63d0\u9ad8\u6570\u636e\u591a\u6837\u6027\u548c\u6a21\u578b\u9c81\u68d2\u6027\u7684\u91cd\u8981\u624b\u6bb5\u3002\u901a\u8fc7\u4f7f\u7528NumPy\u5e93\u751f\u6210\u5404\u79cd\u7c7b\u578b\u7684\u566a\u58f0\uff0c\u5e76\u5c06\u5176\u6dfb\u52a0\u5230\u4e0d\u540c\u7684\u6570\u636e\u96c6\u4e2d\uff0c\u53ef\u4ee5\u6709\u6548\u5730\u589e\u5f3a\u6570\u636e\u96c6\u7684\u591a\u6837\u6027\uff0c\u63d0\u5347\u6a21\u578b\u7684\u6027\u80fd\u3002\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\uff0c\u9700\u8981\u6839\u636e\u5177\u4f53\u573a\u666f\u9009\u62e9\u5408\u9002\u7684\u566a\u58f0\u7c7b\u578b\u548c\u5f3a\u5ea6\uff0c\u4ee5\u5b9e\u73b0\u6700\u4f73\u7684\u589e\u5f3a\u6548\u679c\u3002\u5e0c\u671b\u672c\u6587\u5bf9\u60a8\u5728Python\u7a0b\u5e8f\u4e2d\u52a0\u5165\u566a\u58f0\u6709\u6240\u5e2e\u52a9\u3002<\/p>\n<\/p>\n
\u76f8\u5173\u95ee\u7b54FAQs\uff1a<\/strong><\/h2>\n
 \u5982\u4f55\u5728Python\u4e2d\u4e3a\u6570\u636e\u6dfb\u52a0\u566a\u58f0\uff1f<\/strong>
\u5728Python\u4e2d\uff0c\u53ef\u4ee5\u4f7f\u7528NumPy\u5e93\u751f\u6210\u968f\u673a\u566a\u58f0\u5e76\u5c06\u5176\u6dfb\u52a0\u5230\u6570\u636e\u4e2d\u3002\u5e38\u89c1\u7684\u566a\u58f0\u7c7b\u578b\u5305\u62ec\u9ad8\u65af\u566a\u58f0\u548c\u5747\u5300\u566a\u58f0\u3002\u53ef\u4ee5\u901a\u8fc7numpy.random.normal()<\/code>\u751f\u6210\u9ad8\u65af\u566a\u58f0\uff0c\u901a\u8fc7numpy.random.uniform()<\/code>\u751f\u6210\u5747\u5300\u566a\u58f0\u3002\u5c06\u751f\u6210\u7684\u566a\u58f0\u4e0e\u539f\u59cb\u6570\u636e\u76f8\u52a0\uff0c\u5373\u53ef\u5b9e\u73b0\u566a\u58f0\u7684\u6dfb\u52a0\u3002<\/p>\n
\u6dfb\u52a0\u566a\u58f0\u65f6\u5e94\u8be5\u6ce8\u610f\u54ea\u4e9b\u53c2\u6570\u8bbe\u7f6e\uff1f<\/strong>
\u5728\u6dfb\u52a0\u566a\u58f0\u65f6\uff0c\u5173\u952e\u53c2\u6570\u5305\u62ec\u566a\u58f0\u7684\u5747\u503c\u548c\u65b9\u5dee\uff08\u5bf9\u4e8e\u9ad8\u65af\u566a\u58f0\uff09\u6216\u4e0a\u4e0b\u754c\uff08\u5bf9\u4e8e\u5747\u5300\u566a\u58f0\uff09\u3002\u5747\u503c\u51b3\u5b9a\u4e86\u566a\u58f0\u7684\u4e2d\u5fc3\u4f4d\u7f6e\uff0c\u65b9\u5dee\u5219\u5f71\u54cd\u566a\u58f0\u7684\u5f3a\u5ea6\u3002\u9009\u62e9\u5408\u9002\u7684\u53c2\u6570\u80fd\u591f\u66f4\u597d\u5730\u6a21\u62df\u771f\u5b9e\u4e16\u754c\u4e2d\u7684\u566a\u58f0\u60c5\u51b5\uff0c\u786e\u4fdd\u6570\u636e\u7684\u771f\u5b9e\u6027\u548c\u6709\u6548\u6027\u3002<\/p>\n
\u5982\u4f55\u8bc4\u4f30\u6dfb\u52a0\u566a\u58f0\u540e\u7684\u6570\u636e\u8d28\u91cf\uff1f<\/strong>
\u8bc4\u4f30\u6dfb\u52a0\u566a\u58f0\u540e\u7684\u6570\u636e\u8d28\u91cf\u53ef\u4ee5\u901a\u8fc7\u591a\u79cd\u65b9\u5f0f\u8fdb\u884c\u3002\u53ef\u4ee5\u4f7f\u7528\u53ef\u89c6\u5316\u624b\u6bb5\uff0c\u5982\u7ed8\u5236\u6570\u636e\u5206\u5e03\u56fe\uff0c\u89c2\u5bdf\u566a\u58f0\u5bf9\u6570\u636e\u6a21\u5f0f\u7684\u5f71\u54cd\u3002\u6b64\u5916\uff0c\u8ba1\u7b97\u4fe1\u566a\u6bd4\uff08SNR\uff09\u548c\u5747\u65b9\u8bef\u5dee\uff08MSE\uff09\u4e5f\u662f\u8bc4\u4f30\u6570\u636e\u8d28\u91cf\u7684\u91cd\u8981\u6307\u6807\u3002\u901a\u8fc7\u6bd4\u8f83\u6dfb\u52a0\u566a\u58f0\u524d\u540e\u7684\u6a21\u578b\u6027\u80fd\uff0c\u4e5f\u53ef\u4ee5\u8fdb\u4e00\u6b65\u4e86\u89e3\u566a\u58f0\u5bf9\u6570\u636e\u7684\u5f71\u54cd\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"\u5728Python\u7a0b\u5e8f\u4e2d\u52a0\u5165\u566a\u58f0\u7684\u65b9\u6cd5\u6709\u591a\u79cd\uff0c\u53ef\u4ee5\u901a\u8fc7\u4f7f\u7528NumPy\u5e93\u751f\u6210\u968f\u673a\u566a\u58f0\u3001\u5728\u6570\u636e\u96c6\u4e2d\u6dfb\u52a0\u9ad8\u65af\u566a\u58f0\u3001\u76d0\u548c\u80e1 […]","protected":false},"author":3,"featured_media":1157083,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[37],"tags":[],"acf":[],"_links":{"self":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1157080"}],"collection":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/comments?post=1157080"}],"version-history":[{"count":"1","href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1157080\/revisions"}],"predecessor-version":[{"id":1157084,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1157080\/revisions\/1157084"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media\/1157083"}],"wp:attachment":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media?parent=1157080"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/categories?post=1157080"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/tags?post=1157080"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}